In this project we investigate the mechanisms of DNA replication fidelity in E. coli by a combination of in vivo and in vitro approaches. In vivo, we investigate the specificity of mutation in the E. coli lacI gene in strains affected in various aspects of replication fidelity. For example, analysis of sequenced lacI mutations in wild-type, mismatch repair defective mutL strains, and proofreading defective mutDmutL strains, has allowed estimates to be made for the efficiencies and specificities of in vivo base selection, exonucleolytic proofreading and DNA mismatch repair. In vitro, we have developed novel fidelity assays, again using the lacI gene as a target, allowing measurement of the fidelity of purified DNA polymerase III in its various (sub)assemblies, ranging from the isolated alpha subunit to the complete holoenzyme (HE). Interestingly, the fidelity behavior of polymerase III in vitro is quite different from that in in vivo. Specifically, DNA polymerase III in vitro produces an abnormally high level of (-1) frameshift mutations. This points to the existence of a previously undescribed in vivo fidelity system capable of preventing (-1) frameshifts and other mutations. This system is currently investigated by searching for E. coli mutants defective in this process. We have isolated novel mutants of the dnaX gene (encoding the tau subunit of HE) which specifically enhance frameshifts and transversion mutations, consistent with such a mechanism. We have also developed a system to measure, for the first time, the differences between leading and lagging strand replication on the E. coli chromosome. This system is being exploited to investigate the replication factors that are responsible for this strand-specific fidelity difference.